Many different methods of image guidance are available for radiotherapy treatment (IGRT). The aims of the study were (1) to determine the optimal diameter of gold markers for IGRT to the prostate; (2) to compare, using the Siemens Primatom, the relative merits of in-room computerized tomography (CT) and electronic portal image (EPI) for locating the marker seeds. Gold markers of differing widths were embedded in 2 phantoms (perspex slabs and anthropomorphic). Images were acquired with an amorphous silicon flat panel detector (Siemens Optivue 500) and with the in-room CT scanner (Siemens Somatom Balance). The EPIs were reviewed independently by 6 operators to determine which diameter marker could be best visualized. The optimal marker technique was determined by comparing the investigators' observed marker co-ordinates with the known locations within the phantom. The visibility of all markers on anterior-posterior EPIs was 100%. On the lateral EPI, of a possible 180 visualizations of 1.2-, 1.0-, and 0.8-mm diameter markers, 176 (97.8%), 151 (83.9%), and 132 (73.3%), respectively, were successful. On EPI, the average deviation of fiducial markers from the known position was less than 0.5 mm in any direction. On CT, the largest deviation (2.17 mm) of markers from the known coordinate position was in the superior-inferior direction, reflecting the 3.0-mm slice thickness used. EPI accurately located internal markers in all dimensions. The availability of "gold standard" CT imagery at the treatment unit does not improve how accurately the position of markers in a phantom can be defined compared with EPI. However, CT imagery does provide important soft tissue information, the benefits of which are being investigated further.